What are at-risk customers?

At-risk customers are the ones worth chasing before they are gone: customers whose last order is at least 90 days back and whose lifetime contribution margin exceeds €50. Both halves matter — dormancy alone would flag every one-time bargain hunter, and value alone would flag your best active customers. The intersection is proven value going quiet.

Formula

at risk = days since last order ≥ 90  AND  lifetime margin > €50

Worked example

A store has 7,150 customers; 2,900 haven't ordered in 90 days, but most of those bought once at thin margin and were never coming back. The €50 margin floor cuts the list to 410 — customers who have each earned real profit and have now lapsed. That is a win-back campaign's target list, and 410 × their average LTV is roughly what is walking out the door.

How Saldo Metrics computes it

The widget counts rows of v_customer_lifetime_value where days_since_last_order is 90 or more and lifetime_margin_base exceeds 50 (EUR). Both inputs follow the standard rules: only net-sale orders count, and lifetime margin sums costed orders under the order-grain rule, so a high-value customer whose orders are largely uncosted may sit below the floor until costs land. The same thresholds drive the email win-back recommendations the engine generates, so the KPI, the alert metric and the recommendation feed never disagree about who is at risk.

Why it matters

Win-back is the cheapest revenue a store has — the customer is known, the messaging channel is free, and the margin history says the relationship was profitable. The count trending up is an early retention warning that repeat rate, being a lifetime figure, will only confirm months later.

Common mistakes

• Treating 90 days as churn in slow-cycle categories. A furniture store's customers are all "dormant" between purchases; the fixed threshold fits consumable-to-quarterly purchase rhythms best.
• Blasting the whole list with a percentage discount. These customers were profitable at full price; leading with margin giveaways converts some and retrains the rest to wait for coupons.
• Watching the count without acting on the names. The KPI is a doorbell; the customer-level list behind it is the deliverable.

Where you see this in the app

The At-Risk Customers KPI on the dashboard, the matching alert metric, and the win-back entries in the recommendations feed.